Landslide Susceptibility Zoning in Yunnan Province Based on SBAS-InSAR Technology and a Random Forest Model

Round 1

Reviewer 1 Report (Previous Reviewer 2)

The article has improved much quality compared to the previous version. However, there are still some things that need to be revised, as follows:

1. The section is a part of the methodology. Please consider combining sections 2 and 4.

2. Fonts of the LONG/LAT of the maps in Figures 5,6,7,8,9,10,12,13, and 15 are too big. It needs to be revised.

3. Conclusions should be addressed the main results. It should be removed suggestions or general comments. 

4. The second conclusion should be re-phrased and focus on the main idea as the result of this study because this model is not a new method. 

5. Please check and reduce the similarity rate of the paper.

Author Response

请参阅附件

Author Response File: Author Response.docx

Reviewer 2 Report (Previous Reviewer 3)

Dear Authors, thanks for your response. Still, I believe the manuscript can be improved, and it is not ready for publication. For prediction, we have to have at least two representative models. Gradient Boosting (G-boost) model is more advanced than RFM. Data quantity/size for prediction is also limited.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report (Previous Reviewer 4)

Thank you for submitting your revised paper to Remote Sensing journal. I read the revised manuscript number: remotesensing- 2391670, the manuscript entitled: " Landslide susceptibility zoning in Yunnan Province based on SBAS-InSAR technology and Random Forest Model ". In my point of view, the result of this kind of research could be interesting and useful for many applications specifically for the landslide susceptibility mapping. The authors applied all comments point by point and I confirm their revision. The added information is important and useful and led to improving the manuscript. I accept the revised manuscript in this present form.

Author Response

Thank you for your affirmation of the article, we will definitely continue to work hard to improve the article.

Round 2

Reviewer 2 Report (Previous Reviewer 3)

In this study, RF was used for prediction, therefore, at least 2 methods (XGBoost for example) should be presented and compared. With one method, it is challenging to come up with a decision. Other papers accepted with one method, can not be a reference, yet, when analyzing, at least 2 methods are required. The results of both training and testing should be discussed. Figure 14 also should be illustrated for other cases as well. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

It is of great significance to carry out landslide susceptibility evaluation in Yunnan Province. In this paper, InSAR deformation rate is integrated into the landslide susceptibility evaluation system, which has a good promoting effect on regional landslide susceptibility evaluation.

Author Response

请参阅附件。

Author Response File: Author Response.docx

Reviewer 2 Report

1.      The authors  based on SBAS-InSAR technology and Random Forest Model to study the Landslide susceptibility zoning in Yunnan Province however, we don’t see any information on SBAS-InSAR technology in section 2.

2.      Section 2 should be changed into “Methodology” and present the literature of SBAS-InSAR.

3.      Section 4.2 should be considered moving to section 2 the contents of methodology.

4.      How to define the grading in Table 2?

5.      After dividing into 200x200 what work will you do in the next step?

6.      Section 5 Results and Discussion too short, it should be added more results.

7.      The linking between SBAS-InSAR technology and the random forest model should be addressed.

8.      Please explain why the authors select 5 levels of the susceptibility area? Some references should be cited.

 

9.      Random forest model should be calibrated and verificated.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

My comments are included in the file as attached.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

Reviewer comments

Dear Authors,

Thank you for submitting your paper to the Journal of Sensors. I have carefully reviewed manuscript number Remote Sensing-2271551, "Landslide susceptibility zoning in Yunnan Province based on SBAS-InSAR technology and Random Forest Model". In this paper, the study area chosen was Yunnan Province, where the average annual deformation rate of the radar line-of-sight was obtained for four years (2018-2021) using SBAS-InSAR technology. This rate served as one of the factors in evaluating the susceptibility of Yunnan Province to landslides. The evaluation process involved selecting several index factors, including elevation, slope, aspect, rock group classification, geological structure, rainfall, distance from roads and water systems, topographic undulation, and NDVI. These factors were combined with the annual mean deformation rate, and a random forest model was utilized to accurately analyze and evaluate the risk of landslide geological disasters in Yunnan Province. I believe that the results of this research could be interesting and useful for various applications, such as crisis management. Regarding the English language, I have found that it is moderately written, and I recommend that you carefully check all parts of the manuscript and correct any grammatical errors. Additionally, some sections of the paper require major revisions before any further progress can be made. Please refer to the supplementary comments that I have attached, both in the email and as a PDF file. Thank you again for your submission, and I look forward to seeing the revised version of your manuscript.

Best regards,

Reviewer

 

1- Abstract

1-1- The abstract is well-written and provides a clear overview of the study's objectives, methods, and findings. The language used is technical but understandable, and the key points are succinctly summarized. The abstract includes relevant details about the study area, the technologies used, and the index factors evaluated. It also highlights the importance of the study and its implications for disaster prevention and mitigation. The abstract concludes with a concise summary of the main findings and emphasizes the reliability of the evaluation method used. Overall, the abstract effectively conveys the main points of the paper and provides a good summary for readers. However, it could be improved in the following ways:

·        It would be helpful to provide a brief introduction or context to the problem being addressed in the first sentence.

·        The first sentence is quite long and could be split into two to make it easier to read.

·        It would be helpful to provide some details on the dataset used for the experiments.

·        The conclusion could be more explicit and summarize the contributions of the research.

 

2- Introduction

2-1- In the literature review section, use newer references related to your research from 2020 to 2023.

For Example:

·        Karami, E., Alizadeh, N., Farhadi, H., Abdolazimi, H., and Maghsoudi, Y.: MONITORING OF LAND SURFACE DISPLACEMENT BASED ON SBAS-INSAR TIME-SERIES AND GIS TECHNIQUES: A CASE STUDY OVER THE SHIRAZ METROPOLIS, IRAN, ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., X-4/W1-2022, 371–378, https://doi.org/10.5194/isprs-annals-X-4-W1-2022-371-2023, 2023.

2-2- The current literature review lacks specificity and therefore does not provide sufficient evidence of the novelty of the present study. It would be more effective to expand on the unique contributions of the manuscript in the introduction section. Additionally, the manuscript is not particularly innovative, and it would be helpful to provide further details on its originality.

2-3- The writing structure of the article should be improved.

2-4- Research organization not provided.

 

3- Method

3-1- Research Methodology section were provided in poor way.

3-2- Section "4.2. Data processing" needs related references.

3-3- Flowchart (Figure 1) is not suitable! Must be improved.

3-4- The manuscript is presented as a technical report. Software as a tool is essential, but it should not be the article's focus. Try to focus on contributions and innovations.

3-5- Line 159-163: Long Sentence and lack of references!

3-6- Line 180: " This paper uses DEM data with 30-meter resolution to obtain", Which DEM? Aster? SRTM? or…

3-7- Table 2 is very vague. Explain how to classify indicators. This type of classification makes this work not be implemented correctly in other study areas!

3-8- In Figure 4 (K), the displacement rate is not according to its legend map! what is the reason?

3-9- There have been several studies that have used the Random Forest (RF) technique to assess landslide susceptibility. RF is a machine learning algorithm that can handle large datasets, non-linear relationships, and interactions among variables. It has become a popular method for landslide susceptibility assessment due to its high accuracy and ease of use. One example of a study that used RF for landslide susceptibility mapping was conducted by Arora et al. (2021) in the Indian Himalayas. They collected a range of variables such as slope, aspect, curvature, lithology, and land use/land cover, and generated a landslide inventory map using satellite imagery and field surveys. The authors then used RF to model the relationships between these variables and the landslide occurrences, and generated a susceptibility map. The results showed that RF had high accuracy in predicting landslide occurrences, with an accuracy of 86.2% and an area under the curve (AUC) of 0.89. The study also identified the key variables that contributed to landslide occurrences, with slope, aspect, lithology, and land use/land cover being the most important factors. The authors suggest that their approach can be used for other regions in the Himalayas to generate landslide susceptibility maps. Overall, the RF technique has shown great potential for landslide susceptibility assessment, with several studies reporting high accuracy and robustness in identifying key factors contributing to landslide occurrences. However, it is important to note that the accuracy of the model heavily depends on the quality and availability of the input data, as well as the parameters and settings used for the RF algorithm. With this description, what are your contributions and innovation? Please describe your innovation in full. Otherwise, in my opinion, the article lacks innovation.

 

4 - "Results and Discussion

4-1- This section were provided in poor way, too.

4-2- Improve Figure 5 in terms of dimensions, font and content.

5- The conclusion section needs to rewrite!

 

6- Reference

6-1- The format of the references is not in accordance with the standard of the Remote Sensing journal.

6-2- Would you like to add some new and relevant references?

 

7- There are several problems in the text of the article, some of which are as follows:

Line 125: "source", Check the font

Table 1: "ndvi" must be written "NDVI"

8- Abbreviations should be defined at first mention and used consistently thereafter. For example:

In Line 57: "SAR"

In Line 59: "InSAR"

In Line 143: " SBAS-InSAR"

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The revised version is clearer than the previous version. However, the research design is not so fine.  I think that the paper has some shortcomings regarding content and requires revision, and the authors should address the following items:

1. Most of the figures are not clear, and the text is too big.

2. In section 4, the content on research methods still exists even though there is section 2 (methods).

3. Conclusions too long. Some contents can be moved to the Results and Discussion section. I think the conclusions in the previous version is ok.

 

 

 

 

 

Comments for author File: Comments.pdf

Reviewer 3 Report

The review comments were not addressed sufficiently.

Reviewer 4 Report

    Thank you for submitting your revised paper to Journal of Remote Sensing. I read carefully manuscript number: Remote Sensing-2271551, "Landslide susceptibility zoning in Yunnan Province based on SBAS-InSAR technology and Random Forest Model". Unfortunately, the reviewer's comments have not been fully applied (Point-by-point responses to Reviewer). Please highlight all changes in the revised version of manuscript. Some comments are ignored. I look forward to seeing the revised version of your manuscript (Highlighted).

Yours Sincerely,

Reviewer